KazAnova light is a suite for analytics and data mining 100% written in Java that offers a long range of statistical algorithms and data manipulation tools. It includes all the common modeling techniques as well as it is considerably fast and very user-friendly. It is aimed to assist any person, business, company or educational institution that would like to get a better grasp of credit scoring and data mining and give them the ability to create prediction models in a very short time with high quality, More specifically it includes: 1. Data manipulation tools -Sort -Joins -Aggregations (Group by ) 2. Data pre-processing steps -Optimized binning -Outlier detection -Derivations -Transformations -Partitioning and sampling 3. Statistical and predictability tests -Chi -square (X2) goodness of fit -One Way ANOVA -T-tests -Information value and Weights of Evidence -Information Gain (I-Gain) -Pearson and Spearman Correlations -Basic descriptives -Multicolinearity diagnostics 4. Modeling techniques -Linear (Least square) Regression -Logistic regression -Probit Regression -Decision trees -Neural Networks (Multilayer perception) -Naive Bayes classification -Weights Of evidence (WoE) -Principal components -Discriminant analysis -Nearest Neighbor -Survival Regression 5. Rich graphics' Library -Histograms -Bar Charts -Pie charts -Scatterplots -RoC curves -Kolmogorov-Smirnov 6. Standard reporting tools
Comments not found